Logo Department Name Agency Organization Organization Address Information United States Forest Monongahela National Forest 200 Sycamore Street Department of Service Elkins, WV 26241 Agriculture 304-636-1800

File Code: 1900; 2700 Date: December 23, 2016

Kimberly D. Bose, Secretary Federal Energy Regulatory Commission 888 First St., N.E., Room 1A Washington, DC 20426

Dear Ms. Bose:

Subject: Forest Service Analysis of Landslide Data from a Recent Flood Event on the Monongahela National Forest OEP/DG2E/Gas 4 Atlantic Coast Pipeline, LLC Docket No. CP15-554-000 and CP15-554-001

The Forest Service submits a landslide analysis conducted by Forest Service staff. The analysis utilizes locations of landslides that occurred during the flood event of June 23, 2016 to identify properties of slope, geology, and soil where the landslides occurred. The analysis also compares the results of the properties associated with the landslide analysis with the properties identified along the proposed route of the Atlantic Coast Pipeline (ACP) Project on the Monongahela National Forest. The Forest Service provides this analysis to illustrate the potential for similar high-intensity precipitation events to cause slope stability problems along the proposed ACP route.

Thank you for your attention to this matter. Please direct any questions to Jennifer Adams, Special Project Coordinator, by phone at (540) 265-5114 or by email at [email protected].

Sincerely,

CLYDE THOMPSON Forest Supervisor

cc: Atlantic Coast Pipeline, LLC

Caring for the Land and Serving People Printed on Recycled Paper Landslide Analysis Monongahela National Forest Flood Event (June 2016) December 19, 2016

Introduction Following the June 23, 2016 flood event, heavy rain resulted in numerous landslides across southern West Virginia, some of which occurred within the Monongahela National Forest (MNF). The MNF Engineering staff collected point data (Latitude/Longitude) from 48 landslides. These landslides were found and documented randomly as they were located during road system checks post-flood event. While this point data does not encompass all of the landslides that occurred on the MNF during the flood event, it does offer a “snapshot” of some of the mass movements on forest. As mentioned above, almost all of the landslides used in this analysis were located along road systems. Consequently, it is reasonable to assume that many of these landslides were in part triggered by previous disturbance (i.e. the road system). Although these conditions are not natural, these instances depict how the landscapes in this region respond after they have been disturbed.

The GPS location of these landslides can be used to ascertain environmental conditions present at the area of each landslide. MNF soils staff used the point location of these landslides in conjunction with slope, geology and soils maps to determine the types of environmental conditions present at these landslides. Methods, results and subsequent management considerations of this analysis are described in the paragraphs below.

Objective: Use site specific, fine resolution data (landslide locations) to observe properties where these slides occurred.

Methods Data The GPS coordinates of 48 landslides were used as point data in this analysis. Geospatial data layers including a slope layer created from 3m DEMs, geology map, and soils map were also used for this analysis. The metadata on the slope, geology, and soils layers used in this analysis are summarized in Table 1.

Table1: Metadata on slope, geology, and soil layers used in analysis. Layer Source Scale/Resolution Soil Series NRCS Web Soil Survey 1:24,000 USGS Geology Map USGS 1:24,000 Slope Created from DEMs 3 m

As seen in the table above, the environmental variables used in this analysis are of varying, but relatively coarse scale (as compared to the landslide point data). The MNF acknowledges that there may be some mapping errors as the result of the relatively coarse scale of the environmental data. However, no order 1 soil survey is available for use in this study area, and the environmental variables used in this analysis are of the finest resolution available.

1 Results Geology The geologies where the landslides occurred are listed in Table 2 and summarized below. Attachment A and B display the geology of the northernmost and southernmost recorded landslide points used in this analysis respectively. Attachment C displays the path of the storm.

The Mauch Chunk formation is formed from grayish-red shale, siltstone, sandstone, and some conglomerate. This formation produces soils with shrink-swell clays (3:1:1 vermiculite dominated clays) that are prone to mass movement. The Mauch Chunk formation has the highest potential for slippage of any geology found on the MNF. The Brallier formation is formed from micaceous shale, siltstone, and sandstones. Chemung formations consist of non-calcareous sandstones and shales. The Pottsville formation is predominately gray sandstone and conglomerate which also may contain thin beds of shale, claystone, limestone, and coal. The Pocono group is predominately gray massive sandstones with some shale.

Table 2: Geology of landslides Geology Number of Slides Present on Formation Mauch Chunk 2 Brallier 5 Chemung 13 Pottsville Group 26 Pocono 1 Quaternary Alluvium 1 Total Number of Slides Analyzed: 48

Due to the intensity of the precipitation, landslides occurred on geologies which were previously thought to be relatively stable. The majority of landslides occurred on Pottsville and Chemung as opposed to Mauch Chunk, which is historically known to be at high risk for mass movement. The distribution of landslides should not be interpreted as evidence that landslides are more likely to occur on Pottsville and Chemung than the other geologies. Compared to the other geologies, Pottsville and Chemung covered large portions of the area affected by the stormed (Attachments A and B). However, the fact that many slides occurred on these geologies indicates that they are susceptible to slides during extreme conditions such as those that occurred during the June 23 flood, especially when the soils have been subjected to prior disturbance (prior road construction in this case).

Soil Series Soil series present at identified landslides locations can be seen in Table 3 below. For complete typical characterization of the soil series described below, locate their Official Series Description (OSD) online. The descriptions below focus on the management concerns typical for these soil series. Ernest and Laidig are colluvial soils that can be very deep (> 60 inches) with poor drainage in the subsoil. Shouns series share similar characteristics to those of Ernest and Laidig but are even more susceptible to mass movement due to the fact that Shouns forms from Mauch Chunk geology (and Hampshire Formation) with shrink-swell clay minerals. Potomac soils are very deep and cobbly, floodplain soils. Weikert, Lily and Gilpin are residual soils that are

2 typically very shallow to moderately deep with high rock fragment contents. These soils are typically classified as non-sensitive to earth disturbing activities.

Table 3: Soil series of landslides Soil Series Number of Slides Present on Soil Series Ernest Silt Loam 1 Gilpin-Laidig Association 13 Laidig Channery Silt Loam 4 Lilly Loam 1 Potomac Very Gravelly Loam 1 Shouns Silt Loam 1 Udorthents 10 Weikert Channery Silt Loam 17 Total Number of Slides Analyzed: 48

Based on Table 3, the majority of landslides were recorded on soils that formed from material weathered from interbedded gray and brown acid siltstone, shale, and sandstone. This coincides with the finding from Table 2. As mentioned previously, these soils are not typically associated with management restrictions, but due to the intensity of the storm, these soils became saturated and resulted in mass movement. Due to the poor drainage in the subsoil and the topsoil saturation during the rain event, Laidig also experienced mass movements. This observation also shows that 21% of landslides occurred on Udorthents (human-altered), even though Udorthents typically cover a very small percentage of the overall landscape. These soils have been previously disturbed and as a result are more susceptible to intense rain events due to the instability resulting from altering the natural hydrologic drainage patterns.

Forest Plan standards (SW07) limit the use of wheeled and/or tracked motorized equipment on soil types that are susceptible to landslides—“Use on slopes greater than 15 percent with soils susceptible to downslope movement when loaded, excavated, or wet is allowed only with mitigation measures during periods of freeze-thaw and for one to multiple days following significant rainfall events. If the risk of landslides during these periods cannot be mitigated, then use is prohibited.”

Slope Slope classes present at identified landslides are located in Table 4. The Natural Resource Conservation Service classifies slopes between 20-60% as steep, and anything over 45% classifies as “very steep.” Based on this analysis, the majority of landslides were recorded on slopes of 41-50%. The second most frequent amount of landslides were recorded on 1-10, 51-60, and 61-70% slopes. The likelihood of landslides occurring on each slope class cannot be evaluated because the proportion of the landscape contained in each class is not known. However, the information shows that landslides occurred across all slope classes, including steep slopes, very steep slopes, and slopes that are not classified as steep.

3

Table 4: Slope classes of landslide locations. Slope Range Number of Slides Present on Slope Percent 0-10 6 10-20 5 21-30 2 31-40 4 41-50 17 51-60 6 61-70 6 71-80 1 81-90 1 Grand Total 48

Due to the instability of steep slopes, the Forest Service provides management direction when operating in areas above 40% slope. According to SW07, “Use of wheeled and/or tracked motorized equipment may be limited on soil types include the following soil/site area conditions: Steep Slopes (40 to 50%)—Operation on these slopes shall be analyzed on a case-by-case basis to determine the best method of operation while maintain soil stability and productivity, Very Steep Slope (more than 50%)—Use is prohibited without recommendations from interdisciplinary team review and line officer approval.”

Landslide Summary In summary, the characteristics of the 48 documented landslides are diverse with some common groupings. Many landslides occurred on geologic formations that are normally thought to be relatively stable for mechanized operations (Pottsville and Chemung). However, due to the path of the storm (moving west to east in a relatively straight line across the region), the amount of precipitation in the short duration, and the properties of the landform, the soils even on these more stable landscapes became mobile and mass movement occurred due to soil saturation. There is also evidence in this data to show that previously disturbed soils became susceptible to mass movement (~21% of the slides occurred on Udorthents). Weikert, a shallow soil series (<20 inches) forming on thin shales from the Brailler formation were also highly susceptible to slippage. Landslides occurred across all slope classes, including many on steep slopes as defined by the MNF Forest Plan Standard SW07.

The MNF recognizes that some soil types and geologies in this analysis occur in varying proportions across the landscape. For this reason, we cannot definitively say that one of these soil types or geologies is more susceptible to landslides based on this analysis. However, this analysis does demonstrate that these geologies, even those previously believed to be relatively stable, are highly susceptible to mass movement when steep slopes and extreme weather events coincide. This analysis also demonstrates that disturbed soils (like those near roads) are also prone to mass movement.

4

Impact Footprint Analysis of the Atlantic Coast Pipeline Project Another objective of this analysis was to observe the geology, soil, and slope classes on the current route of the proposed Atlantic Coast Pipeline Project (ACP Project) using the most recent GIS Shapefile (Rev11a) and compare these results with the geology, soil, and slope classes observed in the landslide analysis completed above.

This analysis was intended to bring awareness to the possibility for landslides to occur as a result of intense rain events when certain environmental variables are present (i.e. construction activities on susceptible geologies, soils, and slope percentages).

Results Geology Table 5 presents the results from an analysis completed on the geologic formations within the current ACP Project impact footprint (excluding access roads) on the MNF.

Similar geologies between the landslide analysis and the ACP Project impact footprint analysis occur (both analyses have the presence of Mauch Chunk, Brallier, and Chemung geologies). Based on the observations in Table 5, the majority of the current proposed ACP Project impact footprint includes geologies that are not prone to mass movement under normal circumstances (normal being referred to as non-disturbance activities and/or non-intense rainfall activity). However, as seen in the landslide analysis above (Table 2), certain geologic formations not previously identified as being unstable can experience mass movement under certain intense weather conditions (e.g., Brallier and Chemung).

Table 5: Geologies within the current ACP Project impact footprint on the MNF. Geologic Formations Acres Present Mauch Chunk 13.1 Brallier Formation 20.5 Chemung Group 13.1 Hampshire Formation 25.8 Helderberg Group 0.3 Marcellus Formation of the Millboro Shales 1.7 Mckenzie Formation and Clinton Group 2.2 Oriskany sandstone 1.3 Tonolaway, Wills Creek and Williamsport 1.9 Grand Total 80.0

Soil Series Table 6 presents the results from the soil analysis within the current ACP Project impact footprint (excluding access roads) on the MNF. Shouns silt loam and Weikert channery silt loam are common to both the landslide analysis and the ACP Project footprint.

The majority (~54%) of soils present within the current ACP Project impact footprint is listed as Berks. Berks series is classified as moderately deep and well drained. Weikert soils are residual soils that are typically very shallow to moderately deep with high rock fragment contents. This

5 soil series is typically classified as non-sensitive to earth disturbing activities. Cateache soils are moderately deep and well-drained. This series is typically formed from weathered materials from Mauch Chunk geology. Shouns soil series, as mentioned in the landslide analysis above, is typically formed from material weathered from Mauch Chunk. These soils are comprised of shrink-swell clays that are prone to mass movement. For more information on all of the soil series listed in Table 6, Official Soil Descriptions can be found online.

Table 6: Soil series (USDA NRCS SSURGO 1:24,000 scale) within the current ACP Project impact footprint on the MNF. Soil Series Acres Present Berks Channery Silt Loam 43.14 Calvin-Dekalb-Berks Complex 4.14 Cateache Channery Silt Loam 12.79 Dekalb-Hazleton Complex 1.72 Elleber Extremely Channery Silt Loam 1.57 Shouns Silt Loam 0.30 Weikert Channery Silt Loam 16.37 Grand total 80.04 *The ACP ROW Order 1 Soil Survey has finer scale data that can provide a great geomorphic description of soil properties that result in a higher risk to landslide susceptibility.

Note: An Order 1 Soil Survey was performed for the ACP Project. These results provide finer, more detailed soils information. For the purpose of this analysis, USDA NRCS SSURGO data was used to be consistent with the landslide analysis.

Slope Table 7 presents the results completed from an analysis done on the slope classes present within the ACP Project impact footprint (excluding access roads) on the MNF. From the results listed in Table 7, acres occurring over 40% represent 13% of the current ACP Project impact footprint. The USFS has certain standards that prevent mechanized equipment from operating on slopes greater than 40% due to the increased risk for slope instability.

Table 7: Slope classes within the current ACP Project impact footprint on the MNF. Percent Slope Acres Present 0-10 15.130 10-20 21.563 20-30 19.611 30-40 13.438 40-50 7.785 50-60 2.326 60-70 0.153 70-80 0.027 80-90 0.007 90-100 0.001 Grand Total 80.041

6

*Acres are displayed to .001 decimals so as to show that activities and soil disturbances are proposed on extremely steep slopes and therefore need to be addresses and analyzed.

Conclusions The purpose of this exercise was to identify common geologic formations, soil series, and slope classes present within the current ACP Project impact footprint (excluding access roads) and the landslide points provided by the MNF Engineers that occurred as a result of the flood event on June 23, 2016. Based off of field observations and geospatial data analysis the following conclusions were made:

 Flooding has always affected the Central Appalachian landscape. Storm events, whether related to seasonal weather patterns or tropical systems such as hurricanes, are common within this geographical region. Localized flooding occurs every year. As a result, each year, the MNF faces infrastructure damage and resource impacts. Landslides commonly occur in coves or unstable backslope positions and impact culverts, roads, stream channels, and facilities. Consequently, the flood event from June 2016 should not be considered unusual.  This event and subsequent analysis showed that even more “stable” landforms and geologies are still susceptible to landslides under the right conditions (amount and intensity of precipitation).  Soils that were previously disturbed by management actions can be highly susceptible to mass movement resulting in landslides.  Forest Service standards and guidelines must be followed when operating on steep slopes and soils that are susceptible to slides.

Literature Cited 1Nelson, Stephen A. 2013. Factors that Influence Slope Stability. Tulane University. Accessed at: http://www.tulane.edu/~sanelson/Natural_Disasters/slopestability.htm

Historic Rainfall over Greenbrier County, WV and Alleghany County, VA June 23, 2016. Accessed at: http://www.weather.gov/media/rnk/June23_brief_summary_for_web.pdf

7

ATTACHMENT A

ATTACHMENT B

ATTACHMENT C Historic Rainfall over Greenbrier County WV and Alleghany Co VA June 23, 2016 (Updated July 1, 2016)

The NWS Hydrometeorological Design Studies Center (HDSC) has prepared maps of the Annual Exceedance Probabilities (AEP) of the rainfall in southeast West Virginia and far western Virginia for the June 23-24 event. AEP is often referred to as the Annual Recurrence Interval (e.g. ‘the 100-year event’). The map below shows the AEP for the most intense 24-hour period of the storm, with a ‘thousand-year event in some areas.

Maps of the 1, 3, 6 and 12 hour AEP are also available upon request to NWS Blacksburg. Please visit the HDSC webpage for technical details on AEP: http://www.nws.noaa.gov/oh/hdsc/index.html.

Extremely rare amounts of rainfall on June 23, 2016, most of which occurred in several rounds with high rainfall rates all in a time window of under 12 hours, fell across portions of Greenbrier County West Virginia and Alleghany County Virginia. Radar and rain gauge measurements showed as much as 8-10” (perhaps higher locally) in a band across the central portions of these counties. Return period data suggest this would be nearly a 1 in a thousand year event.

36 hr rainfall ending at 7 or 8am Friday June 24.

Eventually, we will provide more details at a later date, including more of a review of the meteorology, hydrology, and photos of the impacts. What we can say at the time of this writing is that flash flooding and flooding has been historic in these two counties (as well as other counties farther to the northwest in West Virginia), including major to record flood levels on some rivers, numerous water rescues, power outages, structures damaged, and people temporarily displaced. It will be some time before all the details of the impacts are known. Sadly, there are a number of fatalities being reported in Greenbrier County and other parts of West Virginia.